skills.rest
leonardofhy
leonardofhyCOMMUNITY
Explore all skills by this author →
GitHub

hf-research

Coordinate Hugging Face Hub uploads, downloads, searches, and experiment links via CLI.

Updated Aug 27, 2026
Search Tags:#workflow#cli#model#experiment#dataset#huggingface#hf-hub
One-click install
npx skills add https://github.com/leonardofhy/openclaw-workspace --skill hf-research
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: hf-research
Source: https://github.com/leonardofhy/openclaw-workspace/tree/main/skills/hf-research
Command: npx skills add https://github.com/leonardofhy/openclaw-workspace --skill hf-research

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

Hugging Face Hub operations are scattered across tools and manual steps. hf-research provides a unified CLI wrapper that streamlines uploading, downloading, searching, and tracking HF Hub assets, reducing context switching for researchers and engineers.

Core Features & Use Cases

  • Unified interface wrapping hf-cli, hugging-face-model-trainer, and hugging-face-datasets for Hub operations
  • Logs all push/pull actions to memory/hf-research/pushes.jsonl and supports linking results to experiments via EXP-ID
  • Supports status checks, model/dataset discovery, and experiment-result workflows (push-exp) to publish best checkpoints

Quick Start

Run the hf_research CLI to upload, download, search, push experiments, and log HF Hub assets.

Frequently Asked Questions about hf-research

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I automate Hugging Face Hub model and dataset uploads from a single CLI?▼

A unified CLI workflow integrates Hugging Face Hub operations to upload, download, search, and track assets. It wraps the hf-cli, model-trainer, and datasets tools into a single command interface to reduce manual context switching.

How do I push Hugging Face model checkpoints to experiment tracking logs?▼

You can push experiment results by running the push-exp workflow to publish best checkpoints. It logs all push actions to a local jsonl file and links the published results to experiment-manager experiments using an experiment ID.

Can I launch Hugging Face Jobs and track results without manual context switching?▼

Yes, you can launch HF Jobs and discover models directly from the command line. The unified CLI workflow coordinates these operations within a plugin-based workflow to eliminate manual context switching between different tools.

Does the Hugging Face Hub CLI workflow require external dependencies for dataset management?▼

The workflow wraps existing tools like hugging-face-datasets and hugging-face-model-trainer within its plugin-based architecture. It coordinates these tools to manage datasets and models without requiring additional external dependencies for standard Hub operations.

What is the best way to discover and track Hugging Face datasets and models?▼

The best way to discover and track assets is using a unified CLI that searches models and datasets while logging operations. It records all push actions to a local jsonl file and supports linking discovered assets to experiment tracking.

Related Skills

Explore More Software Engineering Skills→
huggingfacehuggingface
10.9k

hf-mcp

Search and retrieve Hugging Face Hub models, datasets, and Spaces via MCP server.

Official
Advanced
gyy0592gyy0592
7

gen-report-detailed

Generate detailed 13-section markdown experiment reports from flame_moonshot training logs and metrics.

Community
Advanced
twjarviszyra-webtwjarviszyra-web
3

weights-and-biases

Track ML experiments, metrics, and artifacts across training runs.

Community
Advanced
LamseyahEliasLamseyahElias

huggingface-hub

Guide hf CLI operations for searching, downloading, and uploading Hugging Face Hub assets.

Community
Intermediate